Se ve ral days after a fire,

Part II of a three-part series on Repair of Fire Damage Evaluating fire damage to concrete structures When the structural condition is ambiguous, a reliable diagnosis can assure future safety and may save money in repairs BY BERNARD ERLIN*, WILLIAM G. HIME*, WILLIAM H. KUENNING * Se ve ral days after a fire, walls and ceilings are cold to both sight and touch. Ash and water or ice lie over eve rything. Co n c rete is likely to be spalled, exposing steel that drapes from beams in shallow arcs. Su rfaces may be blackened, cra c k- ing may be visible and aluminum railings melted. By this time at least a pre l i m i n a ry investigation of the condition of the stru c t u re will perhaps have been made. If the first g e n e ral assessment of the damage has been ambiguous, a re l i a b l e e valuation of the condition of the s t ru c t u re by experts becomes nece s s a ry. There will be great intere s t in this evaluation and considerable u rgency for completing it because i n s u rance companies will be conc e rned about the extent of re p a i r s re q u i red and owners will want the structure back in service with minimum delay. * Erlin, Hime Associates, Materials and Concrete Consultants Northbrook, Illinois ** Consultant, Concrete Technology Lombard, Illinois Spalling of concrete and melting of aluminum railing from gasoline truck fire below a bridge. Although the complete eva l u a- tion will begin with an examination of the condition of va rious materi a l s that went through the fire, such as c e ra m i c s, furn i t u re or metal housi n g s, the information pri m a ri l y needed is the condition of the stru c- t u ral elements that sustain the building loads. Many methods and va rieties of test equipment are ava i l- able for assessing the damage. Some tests can be made at the job site but others re q u i re studies of specimens in the labora t o ry. Although such studies may at times re q u i re sophisticated equipment much or all of the information can often be obtained with quite simple equipment or tools. General survey Su rveys are useful for compiling w h a t e ver information is alre a d y k n own about the nature of the fire. They begin by drawing on inform a- tion in fire department re c o rds and on re p o r ts from witnesses. That i n f o rmation includes the extent and duration of the fire, and the period when failures of members we re observe d. In f o rmation about the type and amount of combustible materi a l s can be used for estimating the pro b-

TABLE I Estimating Temperatures Attained During Fires (Adapted from Reference 1) Factors for Use in Calculating Wood Load Below Multiply weight of material per unit area by the factor indicated to obtain the equivalent wood load. Material Factor Material Factor Material Factor Wood 1.0 Hay or straw 0.8 Cotton 0.9 Cellulose 1.0 Silk 1.2 Paper 0.9 Asphalt 2.1 Wool 1.2 Coal tar 2.1 Coal 1.8 Petroleum products 2.5 Nitrocellulose 0.5 Coke or charcoal 1.5 Lignite 0.8 Animal or vegetable oil 2.1 Rubber 2.1 Peat 1.3 Fats and waxes 2.1 Wood load Equivalent Probable temperature fire test period reached in fire atmosphere Pounds Kilograms (hours) per square feet per square meter Degrees F Degrees C 12.5 61 1 1,700 927 12.5-25 61-122 2 1,850 1,010 25-50 122-244 4 2,000 1,093 TABLE II Conditions of Materials Useful for Estimating Temperature Attained Within a Structure During a Fire (Adapted from Reference 1) Material Typical Examples Condition Degrees F Degrees C Pluming lead; Sharp edges Lead flashing; storage rounded or 550-650 300-350 batteries; toys drops formed Plumbing fixtures; Zinc flashing; galvanized Drops formed 750 400 surfaces Small machine Aluminum parts; brackets; and its alloys toilet fixtures; Drops formed 1,200 650 cooking utensils Softened or adherent 1,300-1,400 700-750 Glass block; jars Molded and bottles; glass tumblers; solid Rounded 1,400 750 ornaments Thoroughly 1,450 800 flowed Window glass; Sheet glass plate glass; Rounded 1,450 800 reinforced glass Thoroughly 1,560 850 flowed Sharp edges Silver Jewelry; rounded or 1,750 950 tableware; coins drops formed Door knobs; Brass furniture knobs; Sharp edges locks; lamp rounded or 1,650-1,850 900-1,000 fixtures; buckles drops formed Sharp edges Bronze Window frames; rounded or 1,850 1,000 art objects drops formed Sharp edges Copper Electric wiring; rounded or 2,000 1,100 coins drops formed Pipes; radiators; Cast iron machine pedestals Drops formed 2,000-2,200 1,100-1,200 and housings able tempera t u res reached within the building. If the building had contained, for example, not much other than about 10 pounds of ve g- etable oil per square foot of floor a rea, all of which was consumed in the fire, probable tempera t u re s reached within the building can be calculated by using the inform a t i o n in Table 1. The amount of ve g e t a b l e oil per square foot multiplied by the factor 2.1 (obtained from the top section of the table) gives a value of 21 pounds per square foot, the e q u i valent wood load to pro d u c e the same amount of heat. In the l ower part of the table it can be seen that a wood load of 21 pounds per s q u a re foot, which falls in the ra n g e of 12.5 to 25 pounds, would supply the amount of heat consumed in a 2-hour fire test. If it is assumed that conditions at the actual fire we re not significantly different fro m those in a standard fire test it would mean that the probable temperat u re reached in the fire atmosphere was 1,850 degrees F (found in the Fa h renheit column). On-site survey Mo re reliable information about t e m p e ra t u res can be obtained during on-site surve y s. One method is to note the condition of materi a l s exposed to the fire (Table 2). Ex a m i- nations of rubble might show that w i n d ow glass had been melted and b rass cupboard knobs had become rounded but that copper in electri c w i res had not been softened. These o b s e rvations would point to temp e ra t u res that had certainly exc e e d- ed 1,560 degrees F (indicated by the glass) and had probably been higher than 1,850 degrees F (indicated by the brass) but not as high as the 2,000 degrees F re q u i red to melt c o p p e r. Those tempera t u re s, howe ve r, might not truly reflect the heat intensity reached in the concre t e, which is likely to be less than in the a i r. Co n c rete color provides a bro a d, g e n e ral guide of tempera t u re s, whether the color re p resents the o riginal surface or one re s u l t i n g f rom spalling. Crazing, cra c k i n g,

Concrete Color Temperature Other Possible Physical Effects Buff 950 C, 1,740 F 1,650 F, 900 C Powdered, light colored, dehydrated paste Black Through Gray to Buff 1,450 F, 800 C Spalling, exposing not more than 25 percent of reinforcing bar surface 600 C 1,100 F 1,070 F, 575 C Popouts over chert or quartz aggregate particles 1,000 F, 550 C Deep cracking Pink to Red 300 C, 550 F 550 F, 300 C Surface crazing Normal 100 F, 40 C None Fig. 1 Visual evidence of temperature to which concrete has been heated (Adapted from References 1 and 2) popouts caused by quartz or chert a g g regate part i c l e s, spalling, and d e h yd ration (crumbling and powd e ring of paste) are general indications of tempera t u res to which conc rete has been exposed, as shown in Fi g u re 1. Co n c rete tempera t u res will va ry from one enclosed area to another and certainly from room to ro o m. The relationship of tempera t u re and strength loss was discussed in the preceding paper of this seri e s. As indicated there, surface temperat u res usually are quite differe n t f rom tempera t u res at differe n t depths in the concre t e, and thus estimates of strength of the intern a l c o n c re t e, when based on surf a c e c o l o r, are speculative. If precise inf o rmation on strength is needed other methods that will be discussed are re q u i re d. The amount and condition of exposed re i n f o rcing bars or pre s t re s s- ing wires and strands are helpful indications for assessing damage. If the steel is draping or shows embri t- tled surf a c e s, serious concrete damage can be anticipated. It does not f o l l ow, howe ve r, that unexposed steel has been as seriously affected, if affected at all. The locations and lengths of exposed steel may be useful for reviews and recalculations of design re q u i re m e n t s. Di s t o rtions and cracking of stru c- t u ral members, as well as their locat i o n s, are useful for evaluating the efficacy of building units for sustaining loads. If cracks are located in c ritical places such as support are a s, if deflections are observed anyw h e re, or if there are misalignments, the support capabilities may have been seriously impaired, whether s u p p o rt is carried by individual members or integrated members that constitute a support system. Equipment commonly used at the site Two ve ry useful tools for examining field concrete are a hammer and chisel. Their resistance to impacts and their resonance when stru c k can re veal information about hardn e s s, integri t y, depth of damage, seriousness of cracking, and the condition of the steel. Other useful tools and test instruments applicable in field studies or in removal of specimens for laborat o ry studies include: impact hamm e r, measuring tape, transit or t h e o d o l i t e, level, dial gages, field m i c ro s c o p e, gri n d e r s, coring machine, wear test devices, steel hardness tester, soniscope, and load test devices.

TABLE III Methods for Details Appraisals of Condition of Fire-Damaged Concrete Condition of Property Methods Notes Actual temperature Examination of building contents See Tables reached in building 1 and 2 Actual temperature Visual examination of concrete, Petrographic, reached in concrete DTA, and metallurgical studies of steel. See Fig. 1 Compressive strength Soundness at highly stressed areas (upper side at center of beam; beam supports; anchorages for reinforcement near support; frame corners) Modulus of elasticity Dehydration of concrete Aggregate performance Spalling Cracking Surface hardness Abrasion resistance Methods for evaluating materials By this time a good idea of the condition of the stru c t u re will have been achieved, and in a few cases enough information will be ava i l- able to close the investigation. Bu t often there will be questions that can be re s o l ved only by detailed s t u d i e s. Methods that can be used to obtain answers to questions re g a rding any of a number of pro p e rties of c o n c rete or the condition of stru c- t u ral members are given in Table 3. Tests on cores. Impact hammer studies. Penetration resistance. Soniscope studies. Hammer and chisel. Visual examination. Soniscope studies. Tests on cores. Soniscope studies. DTA. Petrographic analysis. Chemical analysis. Visual examination. Petrographic analysis. Visual examination.petrographic analysis. Visual examination. Soniscope studies. Petrographic analysis. Dorry hardness or other tests Los Angeles abrasion test on concrete chips* Visual examination for spalling, cracking. Depth of damage Color variation in cores. Chipping. See Fig. 1 Petrographic analysis. Deformation of beams, Gross expansion Differential thermal movements STEEL CONDITION Reinforcing steel, structural steel or prestressing steel Load carrying capacity * Of uncertain value for this purpose. Visual examination. Straightedge and scale. Dial gages or theodolite if needed. Visual examination. Checking of dimensions and levels. Visual check of cores for loss of bond to steel. Color change in concrete next to steel. Physical tests. Metallurgical studies. Dimensional changes, displacement and distortion. Load tests on structure Co res can be used for eva l u a t i n g s t rength and modulus of elasticity. Co res should be taken judiciously and from locations where their effect on strength will be minimal though they provide necessary data. Co m p a risons of these data should be made with data obtained fro m c o res taken from areas that we re not exposed to elevated tempera t u re s. These comparisons provide the most reliable information on changes in concrete caused by the t e m p e ra t u res reached. Co res are also useful for giving incidental information about cracking in the interior of a member, the bond to re i n- f o rcing steel, and interi o r t e m p e ra t u res as re vealed by color c h a n g e s. Some of the inform a t i o n that can be obtained is shown in the schematic drawing of Fi g u re 2. The impact hammer is useful for estimating compre s s i ve stre n g t h, but only when a considerable number of measurements are made and c o m p a red to undamaged concre t e of the same quality from within the same stru c t u re. De t e rminations of pulse ve l o c i- t i e s, using a soniscope, provide indications of strength and modulus, and attenuated signals are indications of cracking. The soniscope is singularly the most effective nondes t ru c t i ve instrument for detecting c ra c k s. Cracking can also be studied visually in cores or fra g m e n t s, but detailed information about cra c k i n g usually must be obtained by using p e t ro g raphic methods. Su i t a b l y p re p a red specimens can re veal mic ro c ra c k s, their relationship to aggregate and steel, and their internal orientation. The extent of concrete dehyd ration can be determined by using p e t ro g raphic methods or by differential thermal analysis (DTA). Sp e c- imens for DTA studies should be c a refully selected, pre f e rably after the petro g raphic studies. The petrog raphic and DTA studies can be extended to assessment of the condition of the aggre g a t e. The hardness and abrasion re s i s- tance of surf a c e s, if they are of interest, can be measured by standard methods and compared to similar data from undamaged concrete in the stru c t u re. Condition of steel Often a major question in determining what stru c t u ral repairs will be needed can be answe red by det e rmining the condition of steel. In floors and walls, exposed pre s t re s s- ing steel can be measured in place using metallurgical methods. If access to the steel is limited because of

height or stru c t u ral conformation of the member, short lengths of steel can be re m oved for labora t o ry studi e s. A micro h a rdness test can be used for determining the tensile s t rength of pre s t ressing steel. This is p a rticularly useful because it is rapid and accurate and re q u i res only a small sample. It correlates we l l with tensile strength determ i n a- t i o n s. Tensile and other tests on pres t ressed or re i n f o rcing steel samples a re valuable because they give a direct measurement of the pro p e rty of g reatest interest. Studies of steel mic ro s t ru c t u re are useful for eva l u a t- ing exposure tempera t u res and their effect on micro s t ru c t u re, and hence on physical pro p e rt i e s. Sonic signal being sent through fire-damaged concrete. Sender and receiver are being held tightly against the concrete; signal is recorded on a test instrument at floor or ground level. Load tests Although a compendium of indirect information about the stru c t u r- al capability of individual elements can be obtained, it may still become n e c e s s a ry, and is indeed fre q u e n t l y d e s i ra b l e, to determine the actual loading capacity of an integra t e d system of elements. For this an actual load test must be perf o rm e d. That test can be made before re i n- statement has been completed, aft e rw a rd, or both. Tests completed p rior to reinstatement will establish what re p a i r s, if any, may be req u i red; tests completed after re i n- statement will indicate whether corrections have been adequate. Us u a l l y, organizations experienced in job-site testing are best qualified for completing the load t e s t s. They are well informed on the va rious methods that can be used, safety precautions that must be emp l oyed, and the interpretation of test data. Fig. 2 Much information can be obtained by petrographic methods from cores taken at appropriate intervals. Data can be used to evaluate the extent, type and severity of damage and to estimate the temperatures to which the concrete was subjected. Serviceability Aside from stru c t u ral considerat i o n s, maintenance and safety of s t ru c t u rally noncritical conditions should be evaluated. For example, a c o n c rete member may be stru c- t u rally adequate, yet careful study of the surface region of the member may re veal incipient spalling or c racking that could result in eve n t u-

al dislodgment of concrete sections, small or larg e. Detailed studies, using petro g raphic methods, may thus be desira b l e. Final evaluation When all of the data are in and b e f o re repairs are undertaken there remain two determinations to be made: (1) what parts of the stru c- t u re cannot be left in their pre s e n t condition? and (2) should corre c- tions be made by re p a i r, re p l a c e- ment, or a combination of the two? These subjects will be discussed in the final paper of this series entitled Reinstating Fi re - Da m a g e d St ru c t u re s. REFERENCES 1. Green, J. K., Reinstatement of Concrete Structures After Fire, (Part 1), The Architects Journal (England), July 14, 1971, pp. 93-99 Photomicrograph shows a major crack, approximately parallel to a formed surface, that outlines an incipient spall caused by thermal shock. The intricate network of fine cracks in the mortar was caused by drying. More cracks were seen by direct microscopic viewing. Petrographic evidence showed temperatures had been high enough to dry the concrete but not to decompose components of the paste or aggregates. 2. Smith, Peter, Investigation of Repair of Damage to Concrete Caused by Formwork and Falsework Fire. ACI Journal, Proceedings V. 60, No. 11, November 1963, pp. 1535-1566. Part I of this three-part article on fire damage to concrete structures appeared in CONCRETE CONSTRUC- TION in March 1972. Part III will appear in May 1972 Core removed for examination. P U B L I C AT I O N# C 7 2 0 1 5 4 Copyright 1972, The Aberdeen Gro u p All rights re s e r v e d